Computing energy distribution

[andicap]

I've seen various methods as to the best way to actually compute Percent Early or Percent Median. I'm wondering what the Sartinistas feel is the correct way.

What I mean is, do you use the raw fractions of the race to compute either the %E or %M?
Do you adjust for a variant and then compute the %E or %M?
Do you adjust for distance changes and/or the daily variant first?
Etc.

P.S. I understand the differences between %E and %M and don't need a treatise on how energy works. I understand all of that. This is purely about mechanics of computing the ratios themselves.

[cj]

I don't see how any raw numbers are going to work very well.

You have huge differences between distances at the same track. Churchill at 6f and 6.5f have vastly different runups that produce much different fractions by horses of similar ability.

You also have huge differences between surfaces. Just compare Churchill 6f races to Pimlico 6f races for one of numerous examples.

[andicap]

I don't see how any raw numbers are going to work very well.

You have huge differences between distances at the same track. Churchill at 6f and 6.5f have vastly different runups that produce much different fractions by horses of similar ability.

You also have huge differences between surfaces. Just compare Churchill 6f races to Pimlico 6f races for one of numerous examples.

I understand --
I'm not looking to debate anyone on this -- I'm just looking for information because different information providers compute %E and %M in different ways and I want to learn exactly why and how they vary.

I especially want to understand how the original Sartin energy ratios were computed. (calling Dick Schmidt.)

For example, some providers make distance adjustments to their times/velocity figures BEFORE they compute the energy ratios. Others compute the track variants and then compute the %E/%M BEFORE they make the adjustments for distance.

[Ted Craven]

Andi,

FWIW, all of Doc Sartin's later programs (from about mid 90's forward) employ the practise of first adjusting raw running times by a Daily Track Variant (DTV), an InterTrack Variant (ITV) and distance equalizing all lines before computing any energy distributions, e.g. fractional velocities, %Median, various factors, Deceleration ratios, etc, etc. This would be true of his software like PaceLauncher, Synthesis, Validator, and later Speculator and RDSS.

I can't compare this series of transformations to an alternate order because I've never attempted it. The foregoing adjustment order seems to have been successful for 10+ years (at least, using DTV and ITV supplied by TrackMaster) so I'm pretty happy with it.

Have you identified some potential advantages to interpolating this order of things?

Ted

[BillW]

I don't see how any raw numbers are going to work very well.

You have huge differences between distances at the same track. Churchill at 6f and 6.5f have vastly different runups that produce much different fractions by horses of similar ability.

You also have huge differences between surfaces. Just compare Churchill 6f races to Pimlico 6f races for one of numerous examples.

Since energy calculations are a ratio does it make a difference?

[46zilzal]

Since energy calculations are a ratio does it make a difference?
Each distance requires a standard, i.e. MOST 7 furlong races run (energy wise) much the same EXCEPT Woodbine which is unique in racing. Know the standards for the specific distance, specific track and go from there.

I never accepted bias until energy distribution numbers showed them to me consistently. I am still amazed.

[cj]

Since energy calculations are a ratio does it make a difference?

Sure, because final times tend not to be affected as much by runup changes. Horses with a long runup get a head start on the clock for quicker pace times, but they tend to tire a little more because the race is longer, distorting the ratio.

[Ted Craven]

Andy,

FYI, here is a 1988 article extracted from Follow Up #12, by Dick Schmidt, with research showing why %Median might be better than %Early. I think it certainly accounts for why Sartin migrated toward %Median after that point.

http://paceandcap.com/forums/showthread.php?t=3426

(post #3)

Ted

[andicap]

thanks for the link, Ted

[Tom]

andicap.....

In the advanced versions of his programs, Sartin used different values for beaten lengths than 10 feet. I don't have the list with me, but it was based on the distance of the race segment - 4 feet, 6 feet, two feet...etc. I'm thinkig it was 12 feet at 6 furlongs, 4 feet at two. Made doifference in the velocity so %M was different as well.

I'll dig them up tonight at home.

[BillW]

Sure, because final times tend not to be affected as much by runup changes. Horses with a long runup get a head start on the clock for quicker pace times, but they tend to tire a little more because the race is longer, distorting the ratio.

Oops, I forgot about that :blush: .

[andicap]

andicap.....

In the advanced versions of his programs, Sartin used different values for beaten lengths than 10 feet. I don't have the list with me, but it was based on the distance of the race segment - 4 feet, 6 feet, two feet...etc. I'm thinkig it was 12 feet at 6 furlongs, 4 feet at two. Made doifference in the velocity so %M was different as well.

I'll dig them up tonight at home.

That gets way too complicated. CJ uses eight feet. If its good enough for him, it's good enough for me.

[46zilzal]

Beaten length difference actually should be calculated depending upon the phase of the contest. It is much more difficult to make up lengths after the 2nd call than to do that earlier. The beaten length adjustment should represent the phase of energy expended.


ON DIRT THAT IS.

[Tom]

That gets way too complicated. CJ uses eight feet. If its good enough for him, it's good enough for me.

Yeah, as long your are consistent. I thought I'd mention it in case you use a Sartin program and things don't match.